Azure HBv3 benchmarks against other Microsoft Azure HPC instance types. Benchmarks by Michael Larabel for a future article on phoronix.
Compare your own system(s) to this result file with the
Phoronix Test Suite by running the command:
phoronix-test-suite benchmark 2104126-PTS-MSAZURE208 Microsoft Azure EPYC 7003 HBv3 Benchmarks - Phoronix Test Suite Microsoft Azure EPYC 7003 HBv3 Benchmarks Azure HBv3 benchmarks against other Microsoft Azure HPC instance types. Benchmarks by Michael Larabel for a future article on phoronix.
HTML result view exported from: https://openbenchmarking.org/result/2104126-PTS-MSAZURE208&grs&sor .
Microsoft Azure EPYC 7003 HBv3 Benchmarks Processor Motherboard Memory Disk Graphics OS Kernel Compiler File-System Screen Resolution System Layer Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 Azure HC - Xeon 8168 2 x AMD EPYC 7V13 64-Core (120 Cores) Microsoft Virtual Machine (Hyper-V UEFI v4.1 BIOS) 442GB 2 x 960GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual Disk hyperv_fb CentOS Linux 8 4.18.0-147.8.1.el8_1.x86_64 (x86_64) GCC 8.3.1 20190507 nfs 1152x864 microsoft 2 x AMD EPYC 7V12 64-Core (120 Cores) Microsoft Virtual Machine (Hyper-V UEFI v4.0 BIOS) 450GB 960GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual Disk 2 x AMD EPYC 7551 32-Core (60 Cores) 226GB 32GB Virtual Disk + 752GB Virtual Disk 2 x Intel Xeon Platinum 8168 (44 Cores) 348GB OpenBenchmarking.org Kernel Details - Transparent Huge Pages: always Compiler Details - --build=x86_64-redhat-linux --disable-libmpx --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-cet --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,fortran,lto --enable-multilib --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-gcc-major-version-only --with-isl --with-linker-hash-style=gnu --with-tune=generic --without-cuda-driver Processor Details - CPU Microcode: 0xffffffff Python Details - Python 3.6.8 Security Details - Azure HBv3 - EPYC 7V13: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full generic retpoline STIBP: disabled RSB filling + tsx_async_abort: Not affected - Azure HBv2 - EPYC 7V12: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full generic retpoline STIBP: disabled RSB filling + tsx_async_abort: Not affected - Azure HBv1 - EPYC 7551: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full generic retpoline STIBP: disabled RSB filling + tsx_async_abort: Not affected - Azure HC - Xeon 8168: itlb_multihit: vulnerable + l1tf: Mitigation of PTE Inversion + mds: Mitigation of Clear buffers; SMT Host state unknown + meltdown: Mitigation of PTI + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full generic retpoline STIBP: disabled RSB filling + tsx_async_abort: Mitigation of Clear buffers; SMT Host state unknown
Microsoft Azure EPYC 7003 HBv3 Benchmarks onednn: Deconvolution Batch shapes_3d - u8s8f32 - CPU onednn: Deconvolution Batch shapes_1d - f32 - CPU pennant: sedovbig compress-zstd: 8 - Compression Speed onednn: Deconvolution Batch shapes_3d - f32 - CPU pennant: leblancbig onednn: Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPU gromacs: Water Benchmark svt-vp9: Visual Quality Optimized - Bosphorus 1080p namd: ATPase Simulation - 327,506 Atoms rodinia: OpenMP LavaMD svt-av1: Enc Mode 8 - 1080p npb: LU.C mnn: inception-v3 botan: ChaCha20Poly1305 botan: ChaCha20Poly1305 - Decrypt lulesh: plaidml: No - Inference - VGG19 - CPU svt-hevc: 1 - Bosphorus 1080p onednn: IP Shapes 1D - f32 - CPU svt-av1: Enc Mode 0 - 1080p plaidml: No - Inference - VGG16 - CPU compress-zstd: 19 - Decompression Speed compress-zstd: 19, Long Mode - Decompression Speed financebench: Bonds OpenMP quantlib: financebench: Repo OpenMP compress-zstd: 8, Long Mode - Decompression Speed rodinia: OpenMP HotSpot3D incompact3d: X3D-benchmarking input.i3d build-nodejs: Time To Compile build-llvm: Time To Compile botan: AES-256 botan: AES-256 - Decrypt compress-zstd: 8, Long Mode - Compression Speed gmpbench: Total Time botan: Twofish botan: Twofish - Decrypt botan: Blowfish - Decrypt botan: Blowfish tnn: CPU - SqueezeNet v1.1 mafft: Multiple Sequence Alignment - LSU RNA botan: CAST-256 botan: KASUMI botan: CAST-256 - Decrypt botan: KASUMI - Decrypt hpcg: build-linux-kernel: Time To Compile plaidml: No - Inference - ResNet 50 - CPU mnn: resnet-v2-50 mnn: SqueezeNetV1.0 tensorflow-lite: SqueezeNet onednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPU onednn: Recurrent Neural Network Training - bf16bf16bf16 - CPU onednn: Matrix Multiply Batch Shapes Transformer - f32 - CPU onednn: Recurrent Neural Network Inference - u8s8f32 - CPU onednn: Recurrent Neural Network Training - u8s8f32 - CPU onednn: Recurrent Neural Network Inference - f32 - CPU onednn: Recurrent Neural Network Training - f32 - CPU onednn: Deconvolution Batch shapes_1d - u8s8f32 - CPU onednn: IP Shapes 3D - u8s8f32 - CPU onednn: IP Shapes 3D - f32 - CPU svt-hevc: 10 - Bosphorus 1080p svt-hevc: 7 - Bosphorus 1080p svt-av1: Enc Mode 4 - 1080p compress-zstd: 19 - Compression Speed cloverleaf: Lagrangian-Eulerian Hydrodynamics Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 Azure HC - Xeon 8168 0.455682 9.50445 5.833419 3184.2 1.58182 3.337201 0.445223 9.041 337.71 0.27566 37.803 94.028 56682.82 34.710 667.765 658.370 41636.603 34.53 45.39 0.881307 0.161 38.39 3717.5 3727.5 104884.294271 2299.3 59196.546875 4256.8 82.489 287.599508 111.327 163.760 5412.131 5407.477 771.9 4893.4 346.319 345.776 425.158 424.738 272.614 14.326 133.691 87.480 133.665 84.059 39.0620 42.008 6.27 29.086 8.224 66320.4 540.194 843.103 0.255501 561.773 962.336 530.898 875.166 0.406686 0.383855 3.21204 548.33 378.01 12.275 78.2 16.66 0.732130 12.1051 6.026449 2653.1 1.59557 3.486503 0.524737 8.458 140.59 0.30056 39.198 78.104 53829.34 55.356 615.795 610.861 34803.539 23.12 36.12 1.12484 0.124 25.57 2631.0 2674.2 133719.328125 1725.2 75154.257813 3037.5 107.976 318.905009 118.969 174.743 3934.401 3938.444 588.0 4155.5 280.902 283.257 349.147 348.425 323.615 16.064 113.600 73.175 113.724 71.101 37.2686 44.922 5.60 52.533 14.136 74885.4 796.816 1294.53 0.350347 791.460 1287.03 778.305 1314.00 1.031942 0.441012 9.68321 379.91 166.90 9.512 69.5 23.78 2.73621 7.91569 15.52899 920.1 3.94077 10.33764 1.15465 3.288 129.44 0.70947 90.175 40.146 25304.19 76.799 303.243 300.178 19011.458 15.97 21.12 1.87834 0.076 19.88 2005.4 2019.3 195702.796875 1251.7 108606.502604 2339.6 145.569 455.812378 194.550 283.291 3144.313 3151.228 468.8 3021.3 217.312 218.034 272.733 273.039 420.047 22.067 88.545 57.950 88.555 55.980 30.4314 58.738 4.80 62.446 17.194 95206.2 3423.15 5355.24 1.51271 3373.88 5539.48 3344.14 5524.50 2.40561 1.76355 10.22263 141.93 119.20 4.669 48.5 25.00 0.668641 2.67914 23.87126 1188.2 1.21181 10.52920 0.398005 4.866 299.95 0.52569 94.119 57.377 35.673 633.217 626.919 20846.590 34.66 23.87 0.931073 0.111 41.54 1903.9 1991.3 130250.449219 1812.9 71609.065104 2350.2 115.832 507.295970 189.096 251.744 3155.391 3159.852 649.3 4196.5 293.551 294.381 365.547 367.942 398.404 15.627 116.630 76.149 116.573 74.287 26.0381 55.324 4.91 29.006 8.226 73517.3 458.201 753.356 0.352180 459.576 746.464 462.855 744.355 0.551467 0.474311 3.28160 480.79 256.28 7.621 76.3 20.19 OpenBenchmarking.org
oneDNN Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 0.6156 1.2312 1.8468 2.4624 3.078 SE +/- 0.002442, N = 3 SE +/- 0.006150, N = 15 SE +/- 0.002363, N = 3 SE +/- 0.005596, N = 3 0.455682 0.668641 0.732130 2.736210 MIN: 0.63 MIN: 0.65 MIN: 2.7 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
oneDNN Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 3 6 9 12 15 SE +/- 0.01628, N = 3 SE +/- 0.06125, N = 3 SE +/- 0.09389, N = 3 SE +/- 0.11811, N = 6 2.67914 7.91569 9.50445 12.10510 MIN: 2 MIN: 5.11 MIN: 4.19 MIN: 5.8 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
Pennant Test: sedovbig OpenBenchmarking.org Hydro Cycle Time - Seconds, Fewer Is Better Pennant 1.0.1 Test: sedovbig Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 Azure HC - Xeon 8168 6 12 18 24 30 SE +/- 0.003905, N = 3 SE +/- 0.012548, N = 3 SE +/- 0.085157, N = 3 SE +/- 0.048781, N = 3 5.833419 6.026449 15.528990 23.871260 1. (CXX) g++ options: -fopenmp -fexceptions -pthread -lmpi_cxx -lmpi
Zstd Compression Compression Level: 8 - Compression Speed OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.9 Compression Level: 8 - Compression Speed Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 700 1400 2100 2800 3500 SE +/- 45.10, N = 3 SE +/- 33.03, N = 15 SE +/- 6.93, N = 3 SE +/- 6.06, N = 3 3184.2 2653.1 1188.2 920.1 1. (CC) gcc options: -O3 -pthread -lz -llzma
oneDNN Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 0.8867 1.7734 2.6601 3.5468 4.4335 SE +/- 0.00120, N = 3 SE +/- 0.00509, N = 3 SE +/- 0.01470, N = 3 SE +/- 0.00862, N = 3 1.21181 1.58182 1.59557 3.94077 MIN: 1.18 MIN: 1.49 MIN: 1.5 MIN: 3.87 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
Pennant Test: leblancbig OpenBenchmarking.org Hydro Cycle Time - Seconds, Fewer Is Better Pennant 1.0.1 Test: leblancbig Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 Azure HC - Xeon 8168 3 6 9 12 15 SE +/- 0.017738, N = 3 SE +/- 0.009939, N = 3 SE +/- 0.027196, N = 3 SE +/- 0.037598, N = 3 3.337201 3.486503 10.337640 10.529200 1. (CXX) g++ options: -fopenmp -fexceptions -pthread -lmpi_cxx -lmpi
oneDNN Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 0.2598 0.5196 0.7794 1.0392 1.299 SE +/- 0.004555, N = 15 SE +/- 0.004912, N = 4 SE +/- 0.004967, N = 3 SE +/- 0.003421, N = 3 0.398005 0.445223 0.524737 1.154650 MIN: 0.31 MIN: 0.39 MIN: 0.48 MIN: 1.1 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
GROMACS Water Benchmark OpenBenchmarking.org Ns Per Day, More Is Better GROMACS 2020.3 Water Benchmark Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 3 6 9 12 15 SE +/- 0.009, N = 3 SE +/- 0.009, N = 3 SE +/- 0.003, N = 3 SE +/- 0.025, N = 15 9.041 8.458 4.866 3.288 1. (CXX) g++ options: -O2 -pthread -lrt -lpthread -lm
SVT-VP9 Tuning: Visual Quality Optimized - Input: Bosphorus 1080p OpenBenchmarking.org Frames Per Second, More Is Better SVT-VP9 0.3 Tuning: Visual Quality Optimized - Input: Bosphorus 1080p Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 70 140 210 280 350 SE +/- 4.36, N = 3 SE +/- 4.18, N = 3 SE +/- 0.20, N = 3 SE +/- 1.10, N = 8 337.71 299.95 140.59 129.44 1. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -O2 -pie -rdynamic -lpthread -lrt -lm
NAMD ATPase Simulation - 327,506 Atoms OpenBenchmarking.org days/ns, Fewer Is Better NAMD 2.14 ATPase Simulation - 327,506 Atoms Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 0.1596 0.3192 0.4788 0.6384 0.798 SE +/- 0.00027, N = 3 SE +/- 0.00059, N = 3 SE +/- 0.00126, N = 3 SE +/- 0.00507, N = 3 0.27566 0.30056 0.52569 0.70947
Rodinia Test: OpenMP LavaMD OpenBenchmarking.org Seconds, Fewer Is Better Rodinia 3.1 Test: OpenMP LavaMD Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 Azure HC - Xeon 8168 20 40 60 80 100 SE +/- 0.23, N = 3 SE +/- 0.20, N = 3 SE +/- 0.39, N = 3 SE +/- 0.02, N = 3 37.80 39.20 90.18 94.12 1. (CXX) g++ options: -O2 -lOpenCL
SVT-AV1 Encoder Mode: Enc Mode 8 - Input: 1080p OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 0.8 Encoder Mode: Enc Mode 8 - Input: 1080p Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 20 40 60 80 100 SE +/- 0.34, N = 3 SE +/- 0.65, N = 3 SE +/- 0.24, N = 3 SE +/- 0.54, N = 12 94.03 78.10 57.38 40.15 1. (CXX) g++ options: -O3 -fcommon -fPIE -fPIC -pie
NAS Parallel Benchmarks Test / Class: LU.C OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: LU.C Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 12K 24K 36K 48K 60K SE +/- 428.57, N = 14 SE +/- 21.24, N = 3 SE +/- 114.01, N = 3 56682.82 53829.34 25304.19 1. (F9X) gfortran options: -O3 -march=native -fexceptions -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi
Mobile Neural Network Model: inception-v3 OpenBenchmarking.org ms, Fewer Is Better Mobile Neural Network 1.1.3 Model: inception-v3 Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 20 40 60 80 100 SE +/- 0.24, N = 15 SE +/- 0.32, N = 3 SE +/- 0.95, N = 12 SE +/- 0.96, N = 12 34.71 35.67 55.36 76.80 MIN: 31.09 / MAX: 427.77 MIN: 34.52 / MAX: 128.55 MIN: 47.56 / MAX: 509.74 MIN: 69.85 / MAX: 676.14 1. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -O2 -rdynamic -pthread -ldl
Botan Test: ChaCha20Poly1305 OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: ChaCha20Poly1305 Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 140 280 420 560 700 SE +/- 0.27, N = 3 SE +/- 0.30, N = 3 SE +/- 0.90, N = 3 SE +/- 1.39, N = 3 667.77 633.22 615.80 303.24 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
Botan Test: ChaCha20Poly1305 - Decrypt OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: ChaCha20Poly1305 - Decrypt Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 140 280 420 560 700 SE +/- 0.06, N = 3 SE +/- 0.55, N = 3 SE +/- 0.78, N = 3 SE +/- 1.37, N = 3 658.37 626.92 610.86 300.18 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
LULESH OpenBenchmarking.org z/s, More Is Better LULESH 2.0.3 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 9K 18K 27K 36K 45K SE +/- 476.10, N = 3 SE +/- 44.77, N = 3 SE +/- 25.50, N = 3 SE +/- 81.47, N = 3 41636.60 34803.54 20846.59 19011.46 1. (CXX) g++ options: -O3 -fopenmp -lm -fexceptions -pthread -lmpi_cxx -lmpi
PlaidML FP16: No - Mode: Inference - Network: VGG19 - Device: CPU OpenBenchmarking.org FPS, More Is Better PlaidML FP16: No - Mode: Inference - Network: VGG19 - Device: CPU Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 8 16 24 32 40 SE +/- 0.19, N = 3 SE +/- 0.42, N = 4 SE +/- 0.31, N = 15 SE +/- 0.21, N = 12 34.66 34.53 23.12 15.97
SVT-HEVC Tuning: 1 - Input: Bosphorus 1080p OpenBenchmarking.org Frames Per Second, More Is Better SVT-HEVC 1.5.0 Tuning: 1 - Input: Bosphorus 1080p Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 10 20 30 40 50 SE +/- 0.60, N = 3 SE +/- 0.13, N = 3 SE +/- 0.02, N = 3 SE +/- 0.12, N = 3 45.39 36.12 23.87 21.12 1. (CC) gcc options: -fPIE -fPIC -O2 -O3 -pie -rdynamic -lpthread -lrt
oneDNN Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 0.4226 0.8452 1.2678 1.6904 2.113 SE +/- 0.012358, N = 3 SE +/- 0.008475, N = 3 SE +/- 0.011740, N = 15 SE +/- 0.021017, N = 3 0.881307 0.931073 1.124840 1.878340 MIN: 0.76 MIN: 0.87 MIN: 1.01 MIN: 1.75 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
SVT-AV1 Encoder Mode: Enc Mode 0 - Input: 1080p OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 0.8 Encoder Mode: Enc Mode 0 - Input: 1080p Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 0.0362 0.0724 0.1086 0.1448 0.181 SE +/- 0.000, N = 3 SE +/- 0.001, N = 3 SE +/- 0.000, N = 3 SE +/- 0.000, N = 3 0.161 0.124 0.111 0.076 1. (CXX) g++ options: -O3 -fcommon -fPIE -fPIC -pie
PlaidML FP16: No - Mode: Inference - Network: VGG16 - Device: CPU OpenBenchmarking.org FPS, More Is Better PlaidML FP16: No - Mode: Inference - Network: VGG16 - Device: CPU Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 9 18 27 36 45 SE +/- 0.22, N = 3 SE +/- 0.39, N = 3 SE +/- 0.21, N = 3 SE +/- 0.21, N = 15 41.54 38.39 25.57 19.88
Zstd Compression Compression Level: 19 - Decompression Speed OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.9 Compression Level: 19 - Decompression Speed Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 Azure HC - Xeon 8168 800 1600 2400 3200 4000 SE +/- 6.14, N = 15 SE +/- 1.17, N = 15 SE +/- 0.66, N = 15 SE +/- 2.22, N = 15 3717.5 2631.0 2005.4 1903.9 1. (CC) gcc options: -O3 -pthread -lz -llzma
Zstd Compression Compression Level: 19, Long Mode - Decompression Speed OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.9 Compression Level: 19, Long Mode - Decompression Speed Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 Azure HC - Xeon 8168 800 1600 2400 3200 4000 SE +/- 12.70, N = 15 SE +/- 2.70, N = 15 SE +/- 0.32, N = 12 SE +/- 2.46, N = 15 3727.5 2674.2 2019.3 1991.3 1. (CC) gcc options: -O3 -pthread -lz -llzma
FinanceBench Benchmark: Bonds OpenMP OpenBenchmarking.org ms, Fewer Is Better FinanceBench 2016-07-25 Benchmark: Bonds OpenMP Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 40K 80K 120K 160K 200K SE +/- 454.05, N = 3 SE +/- 1475.96, N = 4 SE +/- 476.22, N = 3 SE +/- 1191.05, N = 3 104884.29 130250.45 133719.33 195702.80 1. (CXX) g++ options: -O3 -march=native -fopenmp
QuantLib OpenBenchmarking.org MFLOPS, More Is Better QuantLib 1.21 Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 500 1000 1500 2000 2500 SE +/- 4.80, N = 3 SE +/- 7.80, N = 3 SE +/- 0.77, N = 3 SE +/- 1.79, N = 3 2299.3 1812.9 1725.2 1251.7 1. (CXX) g++ options: -O3 -march=native -O2 -rdynamic -lboost_timer -lboost_system -lboost_chrono
FinanceBench Benchmark: Repo OpenMP OpenBenchmarking.org ms, Fewer Is Better FinanceBench 2016-07-25 Benchmark: Repo OpenMP Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 20K 40K 60K 80K 100K SE +/- 228.72, N = 3 SE +/- 15.23, N = 3 SE +/- 166.77, N = 3 SE +/- 324.87, N = 3 59196.55 71609.07 75154.26 108606.50 1. (CXX) g++ options: -O3 -march=native -fopenmp
Zstd Compression Compression Level: 8, Long Mode - Decompression Speed OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.9 Compression Level: 8, Long Mode - Decompression Speed Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 900 1800 2700 3600 4500 SE +/- 4.89, N = 7 SE +/- 4.57, N = 3 SE +/- 2.50, N = 3 SE +/- 1.19, N = 15 4256.8 3037.5 2350.2 2339.6 1. (CC) gcc options: -O3 -pthread -lz -llzma
Rodinia Test: OpenMP HotSpot3D OpenBenchmarking.org Seconds, Fewer Is Better Rodinia 3.1 Test: OpenMP HotSpot3D Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 30 60 90 120 150 SE +/- 1.18, N = 15 SE +/- 0.89, N = 3 SE +/- 0.21, N = 3 SE +/- 1.35, N = 3 82.49 107.98 115.83 145.57 1. (CXX) g++ options: -O2 -lOpenCL
Xcompact3d Incompact3d Input: X3D-benchmarking input.i3d OpenBenchmarking.org Seconds, Fewer Is Better Xcompact3d Incompact3d 2021-03-11 Input: X3D-benchmarking input.i3d Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 Azure HC - Xeon 8168 110 220 330 440 550 SE +/- 0.17, N = 3 SE +/- 0.07, N = 3 SE +/- 2.96, N = 3 SE +/- 7.11, N = 3 287.60 318.91 455.81 507.30 1. (F9X) gfortran options: -cpp -O2 -funroll-loops -floop-optimize -fcray-pointer -fbacktrace -fexceptions -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi
Timed Node.js Compilation Time To Compile OpenBenchmarking.org Seconds, Fewer Is Better Timed Node.js Compilation 15.11 Time To Compile Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 40 80 120 160 200 SE +/- 1.09, N = 3 SE +/- 0.83, N = 3 SE +/- 0.40, N = 3 SE +/- 0.16, N = 3 111.33 118.97 189.10 194.55
Timed LLVM Compilation Time To Compile OpenBenchmarking.org Seconds, Fewer Is Better Timed LLVM Compilation 10.0 Time To Compile Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 60 120 180 240 300 SE +/- 1.89, N = 3 SE +/- 1.99, N = 3 SE +/- 2.31, N = 3 SE +/- 3.41, N = 4 163.76 174.74 251.74 283.29
Botan Test: AES-256 OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: AES-256 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 1200 2400 3600 4800 6000 SE +/- 3.73, N = 3 SE +/- 2.68, N = 3 SE +/- 0.18, N = 3 SE +/- 16.79, N = 3 5412.13 3934.40 3155.39 3144.31 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
Botan Test: AES-256 - Decrypt OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: AES-256 - Decrypt Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 1200 2400 3600 4800 6000 SE +/- 7.85, N = 3 SE +/- 3.26, N = 3 SE +/- 0.03, N = 3 SE +/- 1.16, N = 3 5407.48 3938.44 3159.85 3151.23 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
Zstd Compression Compression Level: 8, Long Mode - Compression Speed OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.9 Compression Level: 8, Long Mode - Compression Speed Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 170 340 510 680 850 SE +/- 6.99, N = 7 SE +/- 6.29, N = 3 SE +/- 0.81, N = 3 SE +/- 3.02, N = 15 771.9 649.3 588.0 468.8 1. (CC) gcc options: -O3 -pthread -lz -llzma
GNU GMP GMPbench Total Time OpenBenchmarking.org GMPbench Score, More Is Better GNU GMP GMPbench 6.2.1 Total Time Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 1000 2000 3000 4000 5000 4893.4 4196.5 4155.5 3021.3 1. (CC) gcc options: -O3 -fomit-frame-pointer -lm
Botan Test: Twofish OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: Twofish Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 80 160 240 320 400 SE +/- 1.10, N = 3 SE +/- 0.20, N = 3 SE +/- 0.15, N = 3 SE +/- 2.54, N = 4 346.32 293.55 280.90 217.31 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
Botan Test: Twofish - Decrypt OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: Twofish - Decrypt Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 80 160 240 320 400 SE +/- 0.77, N = 3 SE +/- 0.13, N = 3 SE +/- 0.22, N = 3 SE +/- 2.43, N = 4 345.78 294.38 283.26 218.03 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
Botan Test: Blowfish - Decrypt OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: Blowfish - Decrypt Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 90 180 270 360 450 SE +/- 0.42, N = 3 SE +/- 0.12, N = 3 SE +/- 0.56, N = 3 SE +/- 0.08, N = 3 425.16 365.55 349.15 272.73 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
Botan Test: Blowfish OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: Blowfish Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 90 180 270 360 450 SE +/- 0.47, N = 3 SE +/- 0.10, N = 3 SE +/- 0.42, N = 3 SE +/- 0.10, N = 3 424.74 367.94 348.43 273.04 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
TNN Target: CPU - Model: SqueezeNet v1.1 OpenBenchmarking.org ms, Fewer Is Better TNN 0.2.3 Target: CPU - Model: SqueezeNet v1.1 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 90 180 270 360 450 SE +/- 0.15, N = 3 SE +/- 0.19, N = 3 SE +/- 0.06, N = 3 SE +/- 0.16, N = 3 272.61 323.62 398.40 420.05 MIN: 272.1 / MAX: 273.5 MIN: 322.04 / MAX: 326.05 MIN: 398 / MAX: 399.03 MIN: 418.9 / MAX: 430.09 1. (CXX) g++ options: -fopenmp -pthread -fvisibility=hidden -O2 -rdynamic -ldl
Timed MAFFT Alignment Multiple Sequence Alignment - LSU RNA OpenBenchmarking.org Seconds, Fewer Is Better Timed MAFFT Alignment 7.471 Multiple Sequence Alignment - LSU RNA Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 5 10 15 20 25 SE +/- 0.10, N = 15 SE +/- 0.13, N = 3 SE +/- 0.06, N = 3 SE +/- 0.24, N = 3 14.33 15.63 16.06 22.07 1. (CC) gcc options: -std=c99 -O3 -lm -lpthread
Botan Test: CAST-256 OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: CAST-256 Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 30 60 90 120 150 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.07, N = 3 SE +/- 0.02, N = 3 133.69 116.63 113.60 88.55 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
Botan Test: KASUMI OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: KASUMI Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 20 40 60 80 100 SE +/- 0.04, N = 3 SE +/- 0.03, N = 3 SE +/- 0.07, N = 3 SE +/- 0.00, N = 3 87.48 76.15 73.18 57.95 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
Botan Test: CAST-256 - Decrypt OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: CAST-256 - Decrypt Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 30 60 90 120 150 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 2 133.67 116.57 113.72 88.56 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
Botan Test: KASUMI - Decrypt OpenBenchmarking.org MiB/s, More Is Better Botan 2.17.3 Test: KASUMI - Decrypt Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 20 40 60 80 100 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.04, N = 3 SE +/- 0.00, N = 3 84.06 74.29 71.10 55.98 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
High Performance Conjugate Gradient OpenBenchmarking.org GFLOP/s, More Is Better High Performance Conjugate Gradient 3.1 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 Azure HC - Xeon 8168 9 18 27 36 45 SE +/- 0.06, N = 3 SE +/- 0.06, N = 3 SE +/- 0.17, N = 3 SE +/- 0.01, N = 3 39.06 37.27 30.43 26.04 1. (CXX) g++ options: -O3 -ffast-math -ftree-vectorize -fexceptions -pthread -lmpi_cxx -lmpi
Timed Linux Kernel Compilation Time To Compile OpenBenchmarking.org Seconds, Fewer Is Better Timed Linux Kernel Compilation 5.10.20 Time To Compile Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 13 26 39 52 65 SE +/- 0.58, N = 15 SE +/- 0.58, N = 13 SE +/- 0.60, N = 15 SE +/- 0.45, N = 13 42.01 44.92 55.32 58.74
PlaidML FP16: No - Mode: Inference - Network: ResNet 50 - Device: CPU OpenBenchmarking.org FPS, More Is Better PlaidML FP16: No - Mode: Inference - Network: ResNet 50 - Device: CPU Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 2 4 6 8 10 SE +/- 0.16, N = 9 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 6.27 5.60 4.91 4.80
Mobile Neural Network Model: resnet-v2-50 OpenBenchmarking.org ms, Fewer Is Better Mobile Neural Network 1.1.3 Model: resnet-v2-50 Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 14 28 42 56 70 SE +/- 0.12, N = 3 SE +/- 0.25, N = 15 SE +/- 1.18, N = 12 SE +/- 1.07, N = 12 29.01 29.09 52.53 62.45 MIN: 27.94 / MAX: 76.84 MIN: 25.79 / MAX: 252.29 MIN: 41.27 / MAX: 326.88 MIN: 54.15 / MAX: 188.44 1. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -O2 -rdynamic -pthread -ldl
Mobile Neural Network Model: SqueezeNetV1.0 OpenBenchmarking.org ms, Fewer Is Better Mobile Neural Network 1.1.3 Model: SqueezeNetV1.0 Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 4 8 12 16 20 SE +/- 0.113, N = 15 SE +/- 0.058, N = 3 SE +/- 0.657, N = 12 SE +/- 0.162, N = 12 8.224 8.226 14.136 17.194 MIN: 5.66 / MAX: 54.17 MIN: 7.29 / MAX: 23.89 MIN: 10.98 / MAX: 78.48 MIN: 15.74 / MAX: 54.06 1. (CXX) g++ options: -std=c++11 -O3 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -O2 -rdynamic -pthread -ldl
TensorFlow Lite Model: SqueezeNet OpenBenchmarking.org Microseconds, Fewer Is Better TensorFlow Lite 2020-08-23 Model: SqueezeNet Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 20K 40K 60K 80K 100K SE +/- 2854.14, N = 15 SE +/- 698.77, N = 3 SE +/- 1151.94, N = 15 SE +/- 1151.23, N = 15 66320.4 73517.3 74885.4 95206.2
oneDNN Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 700 1400 2100 2800 3500 SE +/- 2.64, N = 3 SE +/- 8.06, N = 15 SE +/- 6.35, N = 15 SE +/- 87.86, N = 15 458.20 540.19 796.82 3423.15 MIN: 447.82 MIN: 462.7 MIN: 726.18 MIN: 2651.39 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
oneDNN Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 1100 2200 3300 4400 5500 SE +/- 4.09, N = 3 SE +/- 14.29, N = 15 SE +/- 11.10, N = 15 SE +/- 122.64, N = 15 753.36 843.10 1294.53 5355.24 MIN: 736.05 MIN: 711.66 MIN: 1179.19 MIN: 4475.57 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
oneDNN Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 0.3404 0.6808 1.0212 1.3616 1.702 SE +/- 0.000278, N = 3 SE +/- 0.006544, N = 12 SE +/- 0.011574, N = 15 SE +/- 0.088179, N = 15 0.255501 0.350347 0.352180 1.512710 MIN: 0.3 MIN: 0.27 MIN: 0.57 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
oneDNN Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 700 1400 2100 2800 3500 SE +/- 1.52, N = 3 SE +/- 8.80, N = 15 SE +/- 5.73, N = 15 SE +/- 110.58, N = 12 459.58 561.77 791.46 3373.88 MIN: 450.73 MIN: 471.83 MIN: 718.74 MIN: 2418.62 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
oneDNN Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 1200 2400 3600 4800 6000 SE +/- 3.40, N = 3 SE +/- 102.39, N = 14 SE +/- 13.52, N = 15 SE +/- 103.35, N = 12 746.46 962.34 1287.03 5539.48 MIN: 722.13 MIN: 1149.25 MIN: 4475.65 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
oneDNN Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 700 1400 2100 2800 3500 SE +/- 2.05, N = 3 SE +/- 8.62, N = 15 SE +/- 8.13, N = 3 SE +/- 113.65, N = 12 462.86 530.90 778.31 3344.14 MIN: 450.78 MIN: 458.96 MIN: 738.78 MIN: 2662.07 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
oneDNN Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU Azure HC - Xeon 8168 Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 1200 2400 3600 4800 6000 SE +/- 4.60, N = 3 SE +/- 25.25, N = 12 SE +/- 16.54, N = 15 SE +/- 83.42, N = 13 744.36 875.17 1314.00 5524.50 MIN: 727.38 MIN: 1140.59 MIN: 4658.38 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
oneDNN Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 0.5413 1.0826 1.6239 2.1652 2.7065 SE +/- 0.004961, N = 3 SE +/- 0.003751, N = 3 SE +/- 0.030825, N = 15 SE +/- 0.045752, N = 12 0.406686 0.551467 1.031942 2.405610 MIN: 0.51 MIN: 0.83 MIN: 1.96 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
oneDNN Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 0.3968 0.7936 1.1904 1.5872 1.984 SE +/- 0.009558, N = 12 SE +/- 0.003995, N = 7 SE +/- 0.017521, N = 12 SE +/- 0.040927, N = 15 0.383855 0.441012 0.474311 1.763550 MIN: 0.32 MIN: 0.4 MIN: 0.34 MIN: 0.89 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
oneDNN Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 3 6 9 12 15 SE +/- 0.02162, N = 3 SE +/- 0.00454, N = 3 SE +/- 0.28898, N = 15 SE +/- 0.43651, N = 15 3.21204 3.28160 9.68321 10.22263 MIN: 2.58 MIN: 3.2 MIN: 5.56 MIN: 4.1 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -O2 -pie -lpthread -ldl
SVT-HEVC Tuning: 10 - Input: Bosphorus 1080p OpenBenchmarking.org Frames Per Second, More Is Better SVT-HEVC 1.5.0 Tuning: 10 - Input: Bosphorus 1080p Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 120 240 360 480 600 SE +/- 7.76, N = 3 SE +/- 2.25, N = 3 SE +/- 21.86, N = 15 SE +/- 2.68, N = 15 548.33 480.79 379.91 141.93 1. (CC) gcc options: -fPIE -fPIC -O2 -O3 -pie -rdynamic -lpthread -lrt
SVT-HEVC Tuning: 7 - Input: Bosphorus 1080p OpenBenchmarking.org Frames Per Second, More Is Better SVT-HEVC 1.5.0 Tuning: 7 - Input: Bosphorus 1080p Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 80 160 240 320 400 SE +/- 4.22, N = 3 SE +/- 1.61, N = 3 SE +/- 6.53, N = 12 SE +/- 1.15, N = 3 378.01 256.28 166.90 119.20 1. (CC) gcc options: -fPIE -fPIC -O2 -O3 -pie -rdynamic -lpthread -lrt
SVT-AV1 Encoder Mode: Enc Mode 4 - Input: 1080p OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 0.8 Encoder Mode: Enc Mode 4 - Input: 1080p Azure HBv3 - EPYC 7V13 Azure HBv2 - EPYC 7V12 Azure HC - Xeon 8168 Azure HBv1 - EPYC 7551 3 6 9 12 15 SE +/- 0.031, N = 3 SE +/- 0.218, N = 15 SE +/- 0.023, N = 3 SE +/- 0.034, N = 11 12.275 9.512 7.621 4.669 1. (CXX) g++ options: -O3 -fcommon -fPIE -fPIC -pie
Zstd Compression Compression Level: 19 - Compression Speed OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.9 Compression Level: 19 - Compression Speed Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 20 40 60 80 100 SE +/- 0.71, N = 15 SE +/- 1.01, N = 15 SE +/- 1.18, N = 15 SE +/- 0.83, N = 15 78.2 76.3 69.5 48.5 1. (CC) gcc options: -O3 -pthread -lz -llzma
CloverLeaf Lagrangian-Eulerian Hydrodynamics OpenBenchmarking.org Seconds, Fewer Is Better CloverLeaf Lagrangian-Eulerian Hydrodynamics Azure HBv3 - EPYC 7V13 Azure HC - Xeon 8168 Azure HBv2 - EPYC 7V12 Azure HBv1 - EPYC 7551 6 12 18 24 30 SE +/- 0.83, N = 15 SE +/- 0.26, N = 3 SE +/- 0.45, N = 12 SE +/- 0.11, N = 3 16.66 20.19 23.78 25.00 1. (F9X) gfortran options: -O3 -march=native -funroll-loops -fopenmp
Phoronix Test Suite v10.8.4